test: establish policy and enforcement to prevent future ignored tests

[sanity]

Parent Issue

Part of #2021 - fixing ignored integration tests

Problem

Multiple tests were ignored without tracking issues, leading to hidden bugs and technical debt accumulation. We need systematic prevention.

Root cause of parent issue:

• No enforcement preventing #[ignore] without tracking
• No visibility into accumulating ignored tests
• No documented policy for test management
• Pre-commit hooks don't catch ignored tests

Proposed Solution

Implement multi-layered defense to prevent future ignored test accumulation.

Layer 1: Pre-commit Hook Enhancement

Extend existing TODO-MUST-FIX pattern to cover ignored tests:

// TODO-MUST-FIX: This test has channel closed errors
#[tokio::test]
#[ignore = "Channel closed on startup - investigating"]
async fn test_foo() { ... }

The pre-commit hook should block commits with TODO-MUST-FIX comments, forcing developer to:

• Fix the issue immediately, OR
• Create a tracking issue and link it

Layer 2: CI Check for Ignored Tests

Add CI job that fails if ignored tests exist without tracking issue references:

# Check for ignored tests without issue links
grep -r "#\[ignore" tests/ | grep -v "github.com/.*issues/[0-9]"

Allowed patterns:

• #[ignore = "github.com/freenet/freenet-core/issues/123"]
• #[ignore = "#123 - description"]
• #[ignore = "large scale test - run manually"] (with justification)

Blocked patterns:

• #[ignore] (no reason)
• #[ignore = "fix me"] (no tracking)
• #[ignore = "broken"] (no issue link)

Layer 3: Documentation

Update CLAUDE.md and developer docs with clear test management policy:

## Test Management Policy

### Never Ignore Tests Without Tracking

If a test must be temporarily disabled:

1. Create a GitHub issue documenting:
   - What's broken
   - Why it's broken
   - Steps to reproduce
   - Investigation status

2. Link the issue in the ignore annotation:
   ```rust
   #[ignore = "https://github.com/freenet/freenet-core/issues/123"]

3. OR fix the test immediately

Acceptable Ignore Reasons

Only these scenarios justify ignoring tests:

• Performance/Scale: "large scale test - run manually" (with docs)
• With tracking issue: Link to issue investigating the failure
• Temporary during development: Use TODO-MUST-FIX (blocks commit)

Never Acceptable

• #[ignore] with no reason
• #[ignore = "broken"] or #[ignore = "fix me"]
• Ignoring tests to make CI green
• "We'll fix it later" without an issue

### Layer 4: Regular Audits

Add to sprint planning / release checklist:
- Review ignored tests dashboard
- Verify tracking issues are still active
- Close old issues when tests are re-enabled

## Implementation Tasks

- [ ] Enhance pre-commit hook to check for TODO-MUST-FIX in test files
- [ ] Create CI job: `check-ignored-tests.sh` script
- [ ] Update `.github/workflows/ci.yml` to include ignored test check
- [ ] Document policy in `CLAUDE.md`
- [ ] Document policy in `CONTRIBUTING.md` (if it exists)
- [ ] Create ignored tests dashboard (GitHub project board or query)
- [ ] Add to pre-release checklist: "Review ignored tests"

## Example CI Check Script

```bash
#!/bin/bash
# check-ignored-tests.sh

set -e

echo "Checking for ignored tests without tracking issues..."

# Find all ignored tests
IGNORED_TESTS=$(find . -name "*.rs" -type f -exec grep -H "#\[ignore" {} \;)

if [ -z "$IGNORED_TESTS" ]; then
    echo "✓ No ignored tests found"
    exit 0
fi

# Check each ignored test has a tracking reference
VIOLATIONS=0

while IFS= read -r line; do
    # Allow patterns with issue links or documented reasons
    if echo "$line" | grep -qE "(github\.com|issues/[0-9]+|#[0-9]+|large scale|manual)"; then
        continue
    fi
    
    # Allow empty ignores if followed by TODO-MUST-FIX comment above
    # (pre-commit hook will catch these)
    
    # Flag as violation
    echo "❌ Ignored test without tracking: $line"
    VIOLATIONS=$((VIOLATIONS + 1))
done <<< "$IGNORED_TESTS"

if [ $VIOLATIONS -gt 0 ]; then
    echo ""
    echo "Found $VIOLATIONS ignored test(s) without tracking issues"
    echo "Please either:"
    echo "  1. Fix the test and remove #[ignore]"
    echo "  2. Create a tracking issue and reference it: #[ignore = \"#123 - reason\"]"
    echo "  3. For manual-only tests, justify: #[ignore = \"large scale test - run manually\"]"
    exit 1
fi

echo "✓ All ignored tests have proper tracking"

Success Criteria

• Pre-commit hook blocks TODO-MUST-FIX in tests
• CI check fails on untracked ignored tests
• Policy documented in CLAUDE.md
• Policy documented in CONTRIBUTING.md (if exists)
• Check script tested on current codebase
• All existing ignored tests brought into compliance (via parent issue)
• Dashboard/tracking mechanism established

Future Enhancements

• Auto-generate ignored tests report in CI
• Track "time ignored" metric
• Alert on tests ignored > 90 days
• Integration with GitHub Projects for tracking

Impact

Medium priority - prevents recurrence of parent issue. Should be implemented after fixing existing ignored tests to establish clean baseline.

[AI-assisted debugging and comment]

[github-actions]

🤖 Auto-labeled

Applied labels:

• T-tracking (95% confidence)
• E-medium (80% confidence)
• A-developer-xp (95% confidence)
• A-documentation (85% confidence)

Reasoning: This is a meta/task-oriented issue coordinating multiple changes (pre-commit hook, CI job, docs, dashboard) to prevent ignored tests from accumulating, so T-tracking is most appropriate. The work touches developer tooling and testing policy, so A-developer-xp is relevant; it also requires documentation updates, so A-documentation applies. The scope is moderate (hooks, CI, docs, dashboard), so E-medium reflects expected effort.

Previous labels: E-medium, A-developer-xp

If these labels are incorrect, please update them. This helps improve the auto-labeling system.